Résumé : Machine learning (ML) has advanced quickly, particularly throughout the area of health care. The diagnosis of neurodevelopment problems using ML is a very important area of healthcare. Autism spectrum disorder (ASD) is one of the developmental disorders that is growing the fastest globally. The clinical screening tests used to identify autistic symptoms are expensive and time-consuming. But now that ML has been advanced, it's feasible to identify autism early on. Previously, many different techniques have been used in investigations. Still, none of them have produced the anticipated outcomes when it comes to the capacity to predict autistic features utilizing a clinically validated Indian ASD database. Therefore, this study aimed to develop a simple, quick, and inexpensive technique for identifying ASD by using ML. Various machine learning classifiers, including Adaboost (AB), Gradient Boost (GB), Decision Tree (DT), Logistic Regression (LR), Random Forest (RF), Gaussian Naive Bayes (GNB), Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA), K-Nearest Neighbors (KNN), and Support Vector Machine (SVM), were used to develop the autism prediction model. The proposed method was tested with records from the AIIMS Modified INDT-ASD (AMI) database, which were collected through an application developed by AIIMS in Delhi, India. Feature engineering has been applied to make the proposed solution easier than already available solutions. Using the proposed model, we succeeded in predicting ASD using a minimized set of 20 questions rather than the 28 questions presented in AMI with promising accuracy. In a comparative evaluation, SVM emerged as the superior model among others, with 100 $\pm$ 0.05\% accuracy, higher recall by 5.34\%, and improved accuracy by 2.22\%-6.67\% over RF. We have also introduced a web-based solution supporting both Hindi and English.